diff mbox

[C++,Patch/RFC] PR 55951

Message ID 51519FF8.7040500@oracle.com
State New
Headers show

Commit Message

Paolo Carlini March 26, 2013, 1:17 p.m. UTC 
Hi,

On 03/26/2013 01:32 PM, Jason Merrill wrote:
> On 03/26/2013 08:09 AM, Paolo Carlini wrote:
>> check_array_designated_initializer is called by reshape_init* with
>> ce->index a CONST_DECL, not an INTEGER_CST. Thus I wondered if in such
>> cases it's just matter of using integral_constant_value on it, thus
>> something like the below (which passes testing on x86_64-linux).
> Any constant-expression can go there, so I think I'd use 
> cxx_constant_value unconditionally.
Ok. There is the slight complication that cxx_eval_constant_expression 
cannot cope with identifier nodes (as in desig3.C) and I have to do 
something like the below to pass the testsuite. Is it Ok? Alternately I 
could even check for identifier_p at the beginning of 
cxx_eval_constant_expression.

Thanks,
Paolo.

/////////////////////////

Comments

Jason Merrill March 26, 2013, 2:14 p.m. UTC | #1 
On 03/26/2013 09:17 AM, Paolo Carlini wrote:
> +      if (identifier_p (ce->index))
> +	{
> +	  error ("name %qD used in a GNU-style designated "
> +		 "initializer for an array", ce->index);
> +	  return false;
> +	}
> +
> +      tree ce_index = cxx_constant_value (ce->index);
> +
>         /* The parser only allows identifiers as designated
>   	 initializers.  */
>         if (ce->index == error_mark_node)
>   	error ("name used in a GNU-style designated "
>   	       "initializer for an array");

Let's also combine these two instances of the same error.

Jason
Jason Merrill March 26, 2013, 2:15 p.m. UTC | #2 
On 03/26/2013 10:14 AM, Jason Merrill wrote:
> On 03/26/2013 09:17 AM, Paolo Carlini wrote:
>> +      if (identifier_p (ce->index))
>> +    {
>> +      error ("name %qD used in a GNU-style designated "
>> +         "initializer for an array", ce->index);
>> +      return false;
>> +    }
>> +
>> +      tree ce_index = cxx_constant_value (ce->index);
>> +
>>         /* The parser only allows identifiers as designated
>>        initializers.  */
>>         if (ce->index == error_mark_node)
>>       error ("name used in a GNU-style designated "
>>              "initializer for an array");
>
> Let's also combine these two instances of the same error.

Or at any rate move the error_mark_node case above cxx_constant_value as 
well.

Jason
diff mbox

Patch

Index: cp/decl.c
===================================================================
--- cp/decl.c	(revision 197097)
+++ cp/decl.c	(working copy)
@@ -4766,25 +4766,31 @@  check_array_designated_initializer (const construc
   /* Designated initializers for array elements are not supported.  */
   if (ce->index)
     {
+      if (identifier_p (ce->index))
+	{
+	  error ("name %qD used in a GNU-style designated "
+		 "initializer for an array", ce->index);
+	  return false;
+	}
+
+      tree ce_index = cxx_constant_value (ce->index);
+
       /* The parser only allows identifiers as designated
 	 initializers.  */
       if (ce->index == error_mark_node)
 	error ("name used in a GNU-style designated "
 	       "initializer for an array");
-      else if (TREE_CODE (ce->index) == INTEGER_CST)
+      else if (TREE_CODE (ce_index) == INTEGER_CST)
 	{
 	  /* A C99 designator is OK if it matches the current index.  */
-	  if (TREE_INT_CST_LOW (ce->index) == index)
+	  if (TREE_INT_CST_LOW (ce_index) == index)
 	    return true;
 	  else
 	    sorry ("non-trivial designated initializers not supported");
 	}
       else
-	{
-	  gcc_assert (identifier_p (ce->index));
-	  error ("name %qD used in a GNU-style designated "
-		 "initializer for an array", ce->index);
-	}
+	gcc_unreachable ();
+
       return false;
     }
 
Index: testsuite/g++.dg/ext/desig5.C
===================================================================
--- testsuite/g++.dg/ext/desig5.C	(revision 0)
+++ testsuite/g++.dg/ext/desig5.C	(working copy)
@@ -0,0 +1,7 @@ 
+// PR c++/55951
+
+enum { A };
+
+static const char *a[] = {
+  [A] = "a"
+};